You are on page 1of 8

Proceedings of COBEM 2009 20th International Congress of Mechanical Engineering

Copyright © 2009 by ABCM November 15-20, 2009, Gramado, RS, Brazil

AERODYNAMIC AND MODAL ANALYSES OF BLADES FOR SMALL


WIND TURBINES

Jerson Rogério Pinheiro Vaz, jerson@ufpa.br


Dimitri Oliveira e Silva, dimitrioliveiras@hotmail.com
Alexandre Luiz Amarante Mesquita, alexmesq@ufpa.br
Erb Ferreira Lins, erb@ufpa.br
Federal University of Pará, Institute of Technology, Faculty of Mechanical Engineering, Belém-Pará, Brazil

João Tavares Pinho, jtpinho@ufpa.br


Federal University of Pará, Institute of Technology, Faculty of Electrical Engineering, Belém-Pará, Brazil

Abstract. During the last decades the demand for renewable electric energy production has led to an increasing in
research on wind energy technology. This technology involves technical disciplines such as aerodynamics, structural
dynamics, mechanical as well as electrical engineering. As the improvements in these areas it will lead to further cost
reductions, and for the medium term, wind energy will be able to compete with conventional fossil fuel power
generation technology. The usage of this type of energy generation is quite adequate in certain regions of Amazon,
where there is difficulty of the arrival of electric transmission lines. This paper has as primary objective to present the
results obtained in the studies done in 41 airfoil shapes more used commercially for construction of blades for small
wind turbines. According to operation conditions and dimensions of a rotor, an aerodynamic analysis was performed
and chosen the profile that presents the best power coefficient and angle of attack. In the second part of this paper, a
dynamic analysis was made. Through the finite element method, a modal analysis was accomplished and the first four
modes of vibration of the blade chosen in the previous analysis were determined.

Keywords: Aeolian energy, wind turbine, airfoil shapes, mode shapes, modal analysis.

1. INTRODUCTION

The use of wind turbines to generate electricity has advantages as: generation without pollution, no environmental
impacts, fast installation and commissioning capability, low operation and maintenance cost and taking advantage of
using free and renewable energy (Shokrieh and Rafiee, 2006). In remote areas, wind energy can be used for charging
batteries or can be combined with a diesel engine to save fuel whenever wind is available. Some of the drawbacks are:
the noise produced from wind turbine operation and the fact that wind energy can only be produced when nature
supplies sufficient wind (Hansen, 2008).
Wind turbine transforms the kinetic energy in the wind to mechanical energy in a rotor and shaft and finally into
electric energy in a generator. In general, wind turbine systems consist of five physical components: rotor, transmission,
generator, support structure, and control system (Tempel and Molenaar, 2002). The blade of the rotor is the most
important component in a wind turbine which nowadays is designed according to a refined aerodynamic science in
order to capture the maximum energy from the wind flow (Shokrieh and Rafiee, 2006). The efficiency of a rotor is
characterized by its profile (airfoil section) and the corresponding aerodynamic design. In aerodynamic design,
extensive calculations will be necessary in order to determine the blade parameters, such as chord, thickness and twist
distributions and taper, which are matched with the selected airfoil section (Habali and Saleh, 2000).
Nowadays, blades of horizontal-axis wind turbine (HAWT) are completely made of composite materials. Composite
materials satisfy complex design constraints such as lower weight and proper stiffness, while providing good resistance
to the static and fatigue loading (Shokrieh and Rafiee, 2006). Thus, besides the aerodynamic design, attention must be
paid to dynamic analysis in order to increase their reliability. The fatigue estimation of a rotating blade can be useful for
preventing blade breakage, which is a very frequent problem in the use of wind turbine. The estimation of fatigue for a
rotating blade must go through several steps of calculation: among them the calculation of mode shapes and frequencies
and the computation of displacements and stresses acting on the blades (Mahri and Rouabah, 2002). The numerical
computation of the natural modes of vibration of the blade is necessary to prevent the resonance condition in a design
stage and to comparison with experimental results in order to validate the numerical model.
This paper focuses some aspects in a theory of blade design. The work is divided into two parts. The first one
describes a methodology to choose the profile that best fits the wind conditions for a wind turbine that will be built in
the Amazon region. According to operation conditions and dimensions of the rotor, an aerodynamic analysis was
performed and chosen the profile that presents the best power coefficient and angle of attack. In this study 41 profiles
were analyzed. In the second part of this work, a numerical modal analysis is performed using the finite element
method. In this analysis were determined the first four modes of vibration of the blade chosen in the previous analysis.
Proceedings of COBEM 2009 20th International Congress of Mechanical Engineering
Copyright © 2009 by ABCM November 15-20, 2009, Gramado, RS, Brazil

2. DETERMINATION OF THE ROTOR AIRFOIL

The procedure for the specification of blade profile results from the determination of the optimal angle of attack and
the maximum power coefficient of some profiles suitable for the design of rotors of small wind turbines. For the study,
it was used the Blade Element Moment method (BEM) considering the correction of Glauert (1935) and Prandt
(Eggleston and Stoddard, 1987). The empirical model of Glauert corrects the axial induction factor when it reaches
values greater than 0.4, since the method fails for these values. The factor Prandt considers a finite number of blades.
Table 1 shows the profiles studied in this work.

Table 1. Profiles used for the aerodynamic design of the blade.

Aerodynamics Profiles
E387 NACA23012 NACA64-221
S822 NACA63-209 NACA64-421
SD2030 NACA63-412 NACA65-415
FX63-137 NACA63-415 NACA652-415
S834 NACA63-615 NACA65-418
SH3055 NACA63-218 NACA653-418
NACA1412 NACA63-418 NACA65-618
NACA2408 NACA63-618 NACA653-618
NACA2410 NACA63-221 NACA65-221
NACA2412 NACA63-421 NACA65-421
NACA2415 NACA64-412 NACA654-421
NACA4412 NACA64-415 NACA747A315
NACA4415 NACA64-418 NACA747A415
NACA4418 NACA64-618

The Blade Element Moment method was implemented in MATLAB language and the computational simulations
were made for the following operational characteristics and dimensions of the rotor:
• Length of the blade: 1,75 m;
• Hub radius: 0,15m;
• Wind average speed: 3 m/s;
• Rotational speed: 130 rpm

The results obtained for the more efficient profiles are listed in Table2.

Table 2. Profiles more efficient for wind with average speed of 3 m/s.

Aerodynamic Profiles Maximum Power Angle of Attack (o)


Coefficient
NACA644-421 0.4660 6.5918

NACA653-418 0.4610 4.3469

NACA654-421 0.4660 18.4285

The NACA654-421 has a maximum power coefficient (Cp) for an angle of attack of 18,4285o, where the stall is
almost to occur (see Fig.1), causing a reduction in the lift coefficient generated by an airfoil as angle of attack increases.
This situation does not occur with the NACA644-421. The plot of the Fig. 2 shows that for an angle of attack of 4,3469o
(corresponding to the maximum Cp) the lift coefficient is quite far from the beginning of stall, resulting in the
possibility of the rotor experience higher angles of attack. Thus, the NACA644-421 is the profile chosen for the turbine
blade to be manufactured. Figure 3 shows an illustration of this blade.
For the NACA644-421, the Cp remains almost constant for angles of attack between 3o and 6o and the maximum Cp
(46,6%) occurs for an angle around 6,6o (Fig. 4), however, for higher angles of attack the Cp decreases abruptly.
Proceedings of COBEM 2009 20th International Congress of Mechanical Engineering
Copyright © 2009 by ABCM November 15-20, 2009, Gramado, RS, Brazil

Figure 1. Lift coefficient in relation to the angle of attack for the profile NACA 644-421.

Figure 2. Lift coefficient in relation to the angle of attack for the profile NACA 644-421.

Figure 3. 3D blade with the airfoil NACA 644-421.


Proceedings of COBEM 2009 20th International Congress of Mechanical Engineering
Copyright © 2009 by ABCM November 15-20, 2009, Gramado, RS, Brazil

Figure 4. Power coefficient as function of the angle of attack for the profile NACA 644-421.

In fact, the use of angles of attack calculated for a maximum efficiency does not always correspond to the best rotor
system. It is necessary to examine the continuity of the curve established for the power coefficient. The immediate
response (if the angle of attack used is appropriate) is the plot of the power coefficient as a function of wind speeds and
also as a function of tip speed ratio (TSR). When abrupt decays occurs in the curves of Cp, as observed in Figs. 5 and 6
for angle of attack 6o, it means that, probably the designed rotor when experiences higher wind speeds it will lose lift
force. Therefore the Fig. 5 and Fig. 6 show that the angle of 4o is actually the best.

Figure 5. Comparison between curves of Cp as a function of wind speed for different angles of attack.

Figure 6. Comparison between curves of curves of Cp as a function of the TSR for different angles of attack (NACA
644-421).
Proceedings of COBEM 2009 20th International Congress of Mechanical Engineering
Copyright © 2009 by ABCM November 15-20, 2009, Gramado, RS, Brazil

Figs. 7 and 8 show the chord distribution and the pitch angle as a function of the rotor radius.

Figure 7. Chord distribution for wind rotor using the airfoil NACA 644-421.

Figure 8. Pitch angle distribution for wind rotor using the airfoil NACA 644-421.

3. STRUCTURAL DYNAMIC ANALYSIS

A structural dynamic analysis means to predict the loads on a wind turbine throughout its lifetime and hence the
stress in the material can be computed. Once the dynamic stresses are known, it is possible to calculate the fatigue
damage using standard methods such as Palmgren-Miner rule (Hansen, 2008). Fatigue is a very important issue in a
wind turbine design. Several factors expose wind turbine blades to the fatigue phenomena (Shokrieh and Rafiee, 2006):
• Long and flexible structures;
• Vibration in its resonant mode;
• Randomness in the load spectra due to the nature of the wind;
• Continuous operation under different conditions;
• Low maintenance during lifetime.
For design against fatigue, loads must be determined. All of the loads that occur can be categorized as follow
(Shokrieh and Rafiee, 2006):
• Aerodynamic loads on the blade;
• Weight of the blade;
• Annual gust;
• Changes in the wind direction;
• Centrifugal force;
• Force that arise from start/stop angular acceleration;
• Gyroscopic forces due yaw movements;
• Activation of mechanical brake;
• Thermal effect.
Proceedings of COBEM 2009 20th International Congress of Mechanical Engineering
Copyright © 2009 by ABCM November 15-20, 2009, Gramado, RS, Brazil

A necessary tool for accomplish a structural dynamic analysis is a computer simulation of the dynamic system, with
which to study and understand its dynamic behavior and determine how the system should be modified to change that
behavior in favorable directions. Adequate models can and must permit designers to study and improve the system
behavior as though they were running the real turbine (Spera, 1994). For these computer simulations the finite element
method (FEM) is currently widely used. This methodology permits that accurate models of the blade and the whole
system can be obtained.
In order to have an accurate computational model it is necessary validate this model with experimental data. It is
generally accomplished using modal data, i.e., the mode shape and natural frequencies of the numerical (FEM) model
are computed and validate by mode shape and natural frequencies from experimental modal analysis.

3.1. Numerical Modal Analysis

Analytical (or numerical) and experimental procedures for determining mode shapes and frequencies are referred to
as modal analysis. Mode shapes and natural frequencies are determined primarily by the distribution of mass and
stiffness throughout the structure and by its boundary conditions. Rotation can alter the natural frequencies of certain
mode shapes, when centrifugal and Coriolis forces change stiffness (Spera, 1994). A plot that shows the variation of the
natural frequencies as a function of the rotational speed is the Campbell diagram. Figure 8 shows the results of a
Campbell diagram displaying the results of a modal analysis in an HAWT (Yamane et al., 1992). In this plot the thick
solid lines and dotted lines denote natural frequencies, and the radial lines of nP can be regarded as the frequency of the
excitation force (P means the rotational speed and n is a integer multiple of the rotational speed). The resonance is a
phenomenon which occurs in structure when the frequency of the excitation equals or nearly equals one of the modal
frequencies of the structure. Thus, in Fig. 9, every intersection of a radial line (nP) and a modal frequency line is a
potential resonance. However, in order to validate models is more common accomplish the numerical and experimental
modal analysis for the structure without rotation. Therefore, the numerical modal analysis for the blade of this work
does not take into account the rotational speed of the rotor. This analysis is shown in the next section.

Figure 9. Campbell diagram of a wind turbine (Yamane et al., 1992).

3.2. Results of Modes Shapes and Frequencies of the Blade with Profile NACA 644-421

Once defined the blade profile for the wind turbine that will be built in a site in Amazon, the next step is to create
the finite element model of the blade to obtain the modal parameters (mode shapes and natural frequencies). As the
rotor is very short, the blade will be manufactured in a peculiar manner: first is created a mold of the blade according to
its length and profile. After that, this mold is filled with polyurethane, and hence, it is inserted an aluminum tube inside
of the mold. This tube will make the blade stronger and stiffer and will also have the function of connecting the blade to
the hub. After drying the mixture, the polyurethane and the tube are taken off of the mold; hence a thin layer of glass
fiber is placed externally to the polyurethane in order to give more strength to the blade.
For the analysis in ANSYS, the material properties of the blade are listed in Table 3. The geometric data of the blade
are shown in Table 4. The tube of aluminum has 0.020 m of external diameter and 0.016 of internal diameter.
Proceedings of COBEM 2009 20th International Congress of Mechanical Engineering
Copyright © 2009 by ABCM November 15-20, 2009, Gramado, RS, Brazil

Table 3 – Material properties of the blade.

Material Aluminium Polyurethane Glass fiber


Density 2700 kg/m³ 1150 Kg/m³ 2440 Kg/m³
Modulus of elasticity 70 Gpa 1.72 Mpa 68.9 Gpa
Coefficient of Poisson 0.35 0.103 0.183

Table 4 – Dimensions of the blade.

Profile NACA 644-421


Length of the blade 1.75 m
Maximum chord 0.26 m
Minimum chord 0.05 m

The characteristics of the pre-processing in the analysis are:


• Finite element: Solid 45;
• Mesh: 268459 elements;
• Mode shape extractor method: Block Lanczos;
• Boundary conditions: Fixed-free. The aluminum tube is fixed to the hub.

The natural frequencies computed are listed in Table 5 and the mode shapes are shown in Fig. 9.

Table 5 – Natural frequencies.

1st Mode 2nd Mode 3rd Mode 4th Mode


Frequency [Hz] 3.863 3.941 11.885 38.162

Figure 9. Mode shapes of the blade.


Proceedings of COBEM 2009 20th International Congress of Mechanical Engineering
Copyright © 2009 by ABCM November 15-20, 2009, Gramado, RS, Brazil

4. CONCLUDING REMARKS

This paper presented an aerodynamic and a dynamic analysis of a blade for a wind turbine. This turbine will be
constructed in a site in Amazon where does not exist connection to the national grid of electric energy. In aerodynamic
analysis, it was determined a profile that results in better power coefficient and angle of attack. In the numerical modal
analysis, through finite element method, the four first natural modes of vibration were calculated. It was observed that
the first modal frequency (231 rpm) is higher than the nominal rotational speed of the turbine (130 rpm), thus avoiding
resonance.

5. ACKNOWLEDGMENTS

The authors would like to thank the CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) for
the financial support of this project.

6. REFERENCES

Eggleston, D. M., Stoddard, F. S., 1987, “Wind Turbine Enginering Design”, Van Nostrand Reinhold Company, New
York.
Glauert, H., 1935, “AirPlane Propellers, in Aerodynamic Theory”, Spring Verlag, ed. W. F. Durand, pp. 169-360.
Habali, S. M., Saleh, I. A., 2000, “Local design, testing and manufacturing of small mixed airfoil wind turbine blades of
glass fiber reinforced plastics. Part I: Design of the blade and root”, Energy Conversion & Management, 41, pp 249-
280
Hansen, M., 2008, “Aerodynamics of Wind Turbines”, 2nd Edition, Earthscan.
Mahri, Z. L., Roubah, M.S., 2002, “Fatigue Estimation for a Rotating Blade of a Wind Turbine”. Rev. Energ. Ren.,
Vol.5, pp39-47
Shokrieh, M. M., Rafiee, R., 2006, “Simulation of fatigue failure in a full composite wind turbine blade”, Composite
Structures, 74, pp 332–342
Spera, D. , 1994, “Wind Turbine Technology”, ASME Press.
Tempel, J.V., Molenaar, D.P., 2002,“Wind Turbine Structural Dynamics –A Review of the Principles for Modern
Power Generation, Onshore and Offshore”, Wind Engineering, Vol 26, NO. 4 , pp 211–220
Yamane, T., Matsumiya, H., Kawamura, S., Mizutani, H., Nii, Y, Gotanda, T., 1992,“Vibration Characteristics of an
Experimental Wind Turbine with a 15 m Teetered Rotor”, The Japan Society of Mechanical Engineers, Vol. 35, No
02, pp. 268-273

7. RESPONSIBILITY NOTICE

The authors are the only responsible for the printed material included in this paper.